Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

Elon Musk's updated Grok AI claims to be better at coding and math

Mar 29, 2024 - engadget.com
Elon Musk's xAI has launched an update to its AI, Grok-1.5, with improved capabilities in math, coding, and reasoning, and the ability to process longer contexts. The company claims that Grok-1.5 now competes with GPT-4, Gemini Pro 1.5, and Claude 3 Opus in several areas. The AI's performance has significantly improved, with a 50.6% score in the MATH benchmark, and 90% and 74.1% in GSM8K (math word problems) and HumanEval (coding) respectively. It can also process long contexts of up to 128K tokens within its context window, allowing it to utilize information from longer documents.

However, xAI did not provide details on Grok's progress in other areas, such as academic scores and multimodal. Grok-1.5's position may not be secure for long, as ChatGPT 5 is set to launch this summer with a feature set that promises a more human-like communication experience. Currently, Grok is only available for Premium+ tier users on X (formerly Twitter), but Musk plans to make it available to regular Premium users. The company has also open-sourced its Grok chatbot.

Key takeaways:

  • Elon Musk's xAI has launched Grok-1.5, an update to its AI, with improved capabilities in math, coding and the ability to process longer contexts.
  • The company claims that Grok-1.5 now competes with GPT-4, Gemini Pro 1.5 and Claude 3 Opus, with significant improvements in MATH benchmark, GSM8K and HumanEval scores.
  • Grok-1.5 can process long contexts of up to 128K tokens within its context window, allowing it to utilize information from substantially longer documents.
  • Currently, Grok is only available for users of the Premium+ tier on X (formerly Twitter), but Elon Musk has promised to open it up to X's regular Premium users.
View Full Article

Comments (0)

Be the first to comment!